LCDS: A Logic-Controlled Discharge Summary Generation System Supporting Source Attribution and Expert Review
Cheng Yuan, Xinkai Rui, Yongqi Fan, Yawei Fan, Boyang Zhong, Jiacheng Wang, Weiyan Zhang, Tong Ruan

TL;DR
LCDS is a system that improves discharge summary generation by constraining content scope, ensuring source attribution, and supporting expert review to reduce hallucinations in LLM outputs.
Contribution
It introduces a logic-controlled framework with source mapping and attribution, enhancing reliability and reviewability of discharge summaries generated by LLMs.
Findings
Enhanced source attribution for generated summaries
Reduced hallucination issues in discharge summaries
Facilitated expert review and iterative improvement
Abstract
Despite the remarkable performance of Large Language Models (LLMs) in automated discharge summary generation, they still suffer from hallucination issues, such as generating inaccurate content or fabricating information without valid sources. In addition, electronic medical records (EMRs) typically consist of long-form data, making it challenging for LLMs to attribute the generated content to the sources. To address these challenges, we propose LCDS, a Logic-Controlled Discharge Summary generation system. LCDS constructs a source mapping table by calculating textual similarity between EMRs and discharge summaries to constrain the scope of summarized content. Moreover, LCDS incorporates a comprehensive set of logical rules, enabling it to generate more reliable silver discharge summaries tailored to different clinical fields. Furthermore, LCDS supports source attribution for generated…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
Taxonomy
TopicsTopic Modeling · Machine Learning in Healthcare · Text Readability and Simplification
